US5589409A

US5589409A - Fabrication of bipolar transistors with improved output current-voltage characteristics

Info

Publication number: US5589409A
Application number: US08/393,647
Authority: US
Inventors: Constantin Bulucea; Michael J. Grubisich
Original assignee: National Semiconductor Corp
Current assignee: National Semiconductor Corp
Priority date: 1994-09-02
Filing date: 1995-02-24
Publication date: 1996-12-31
Anticipated expiration: 2014-09-02
Also published as: US5548158A; DE69509211T2; DE69509211D1; EP0727097A1; EP0727097B1; WO1996008039A1

Abstract

A special two-dimensional intrinsic base doping profile is utilized to improve the output current-voltage characteristics of a vertical bipolar transistor whose intrinsic base includes a main intrinsic portion. The special doping profile is achieved with a pair of more lightly doped base portions that encroach substantially into the intrinsic base below the main intrinsic base portion. The two deep encroaching base portions extend sufficiently close to each other to set up a two-dimensional charge-sharing mechanism that typically raises the magnitude of the punch-through voltage. The transistor's current-voltage characteristics are thereby enhanced. Manufacture of the transistor entails introducing suitable dopants into a semiconductor body. In one fabrication process, a fast-diffusing dopant is employed in forming the deep encroaching base portions without significantly affecting earlier-created transistor regions.

Description

This is a division of U.S. patent application Ser. No. 08/300,498, filed 2 Sep. 1994.

FIELD OF USE

This invention relates to semiconductor devices. More particularly, this invention relates to bipolar transistors, including structures and methods for fabricating bipolar transistors.

BACKGROUND

A bipolar transistor formed with an emitter, collector, and intervening base is typically created in a vertical arrangement along a major surface, referred to here (for convenience) as the upper surface, of a monocrystalline semiconductor body. The emitter, collector, and base are all situated in the semiconductor body. The emitter adjoins the body's upper surface.

The base consists of an intrinsic part (commonly referred to as the "intrinsic base") and one or more laterally adjoining extrinsic parts (commonly referred to as "extrinsic bases"). The intrinsic base lies directly below the emitter. In a standard symmetrical configuration, there are two extrinsic bases located on opposite sides of the intrinsic base. Each extrinsic base includes a heavily doped contact zone which extends to the upper semiconductor surface and to which electrical contact is made at a location spaced laterally apart from the emitter. Each extrinsic base may include a connection (or link) zone that provides a low-resistance electrical path between the intrinsic base and the associated base contact zone.

The collector contains a buried main portion situated below the intrinsic base so that the emitter-collector current flows generally in the vertical direction. The main collector portion extends laterally beyond the intrinsic base to a heavily doped collector contact zone which typically extends to the upper semiconductor surface to provide electrical access to the collector at a location spaced laterally apart from the emitter and the two base contact zones. Overlying ohmic electrical contacts to the emitter and contact zones complete the transistor.

Referring to the drawings, FIG. 1 illustrates a prior art oxide-isolated npn transistor of the vertical type as described in De Jong et al, U.S. Pat. No. 5,006,476. The semiconductor body in FIG. 1 consists of p- monocrystalline silicon substrate 10 and overlying n- epitaxial silicon layer 12. Recessed field-oxide region 14 divides the epitaxial silicon into a group of laterally separated active semiconductor regions, two of which are interconnected by n+ buried layer 16 situated along the metallurgical interface between substrate 10 and epitaxial layer 12.

In the device of FIG. 1, n+ emitter 18, which is formed by dopant outdiffusion from n+ polycrystalline silicon ("polysilicon") emitter contact 20 so as to be self-aligned to emitter contact 20, overlies p intrinsic base 22. The transistor has two extrinsic bases, each formed with a p+ contact zone 24 and a p connection zone 26. N+ collector contact zone 28 connects to a main collector portion consisting of n+ buried layer 16 and overlying n-type

epitaxial portions

30 and 32.

Metal silicide caps

34, 36, and 38 respectively overlie

contact zones

20, 24, and 28. Oxide spacers 40 provide spacing between emitter contact 20 and each metal silicide cap 36 through which a metal line (not shown) is connected to underlying base contact zone 24.

Numerous engineering trade-offs are made in designing vertical bipolar transistors. For example, by incorporating spacers 40 into the npn transistor of FIG. 1, the probability of an electrical short between emitter contact 20 and either of the base contacts is typically quite low. Base connection zones 26 accompany spacers 40 in order to provide low-resistance links between intrinsic base 22 and each base contact zone 24. However, forming connection zones 26 typically entails additional processing and thus additional manufacturing expense.

As the lateral dimensions of transistors shrink, the probability that heavily doped base contact zones 24 will encroach on emitter 22 increases. Tang et al, "Design Considerations of High Performance Narrow-Emitter Bipolar Transistors, " IEEE Elect. Dev. Lett., April 1987, pp. 174-175, and Lu, "Lateral Encroachment of Extrinsic-Base Dopant in Submicrometer Bipolar Transistors," IEEE Elect. Dev. Lett., October 1987, pp. 496-498 both investigate the effect of such encroachment in vertical bipolar transistors similar to that of FIG. 1. Tang et al reports decreases in the transistor cutoff frequency f_T and the small-signal transistor current gain. Cutoff frequency f_T is the frequency at which the small-signal current gain drops to 1. Lu et al reports a decrease in the collector current I_c.

An important parameter in bipolar transistor design is the punch-through voltage V_P. This is the value of the collector-to-emitter voltage V_CE at which the depletion region of the collector-base junction reaches the depletion region of the emitter-base junction when the base is open or the when base-to-emitter voltage V_BE is at a selected value.

FIGS. 2a, 2b, 3a, and 3b are helpful in understanding the punch-through phenomenon. FIG. 2a illustrates a simplified one-dimensional version of an npn transistor during normal transistor operation, where t_E and t_B respectively are the emitter and base thicknesses. FIG. 3a illustrates the transistor when it is in the punch-through condition. FIGS. 2b and 3b respectively depict the potentials across the transistor when it is in the conditions of FIGS. 2a and 3a.

During normal operation, electrons in the emitter must overcome the potential barrier at the emitter-base junction in order to enter the quasi-neutral region of the base between the two depletion regions. Electrons then diffuse across the quasi-neutral region until they reach the collector-base depletion region where the electric field pulls them into the collector. See FIGS. 2a and 2b. Altogether, the current flow from emitter to collector is barrier limited and diffusion limited. The magnitude of collector-to-emitter voltage V_CE does not substantially impact the magnitude of the emitter-to-collector current during normal operation because (a) the potential barrier at the emitter-base junction is controlled by the base-to-emitter voltage V_BE and (b) the diffusion-limiting action on the emitter-to-collector current flow is determined by the thickness of the quasi-neutral region of the base.

When collector-to-emitter voltage V_CE is raised to such a value that the collector-base depletion region punches through to the emitter-base depletion region, the quasi-neutral region of the base is eliminated as shown in FIGS. 3a and 3b. The diffusion limitation on the emitter-to-collector current flow is no longer present. Also, voltage V_CE now directly influences the emitter-base barrier potential, causing its magnitude to be easily reduced. The net result is that the number of electrons passing through the base increases rapidly in a generally undesirable manner as voltage V_CE is increased. Accordingly, it is desirable that the magnitude of punch-through voltage V_P be high.

Another important bipolar design parameter is the Early voltage V_A. At a given value of base-emitter voltage V_BE, collector current I_C is ideally independent of collector-to-emitter voltage V_CE across the normal V_C operating range. However, because the size of the quasi-neutral region of the base decreases with increasing V_CE, I_C actually increases slowly with increasing V_CE at fixed V_BE in the manner generally shown in FIG. 4.

Early voltage V_A is a transistor modeling parameter that constitutes the voltage at which the collector current asymptotes approximately intersect the V_CE axis. The points at which the I_C asymptotes at various V_BE values intersect the V_CE axis do not exactly coincide in a real bipolar transistor. Nonetheless, it has turned out to be a good modeling parameter to consider the I_C asymptotes as all meeting at the V_A position on the V_CE axis. As with punch-through voltage V_P, it is desirable that the magnitude of the Early voltage V_A be high, ideally infinite.

A critical engineering trade-off relating to the intrinsic base involves punch-through voltage V_P, Early voltage V_A, and the base (series) resistance r_B, on one hand, and cutoff frequency f_T and collector current I_C, on the other hand. Reducing base thickness t_B and/or the total base doping causes f_T and I_C to increase. The current gain also increases with reduced t_B. However, V_P, V_A, and r_B are degraded when t_B and/or the base doping is reduced.

In a simplified one-dimensional analysis, this trade-off can be seen from the following equations that apply to a symmetrical single-emitter vertical npn transistor such as that shown in FIG. 1: ##EQU1## where: x is an integrating variable in the base along the direction of main current flow,

α₀ is the static common-base current gain (nearly 1),

D_n is the average electron diffusivity in the base,

t_B is the metallurgical thickness of the base--i.e., the distance between the emitter-base and collector-base junctions,

t_BEFF is the effective electrical thickness of the base--i.e., the distance between the boundaries of the emitter-base and collector-base depletion regions,

n_i is the intrinsic electron density (approximately 1.4×10¹⁰ electrons/cm³ in silicon at room temperature),

N_A is the base (acceptor) dopant concentration,

W_E is the lateral width of the emitter,

L_E is the length of the emitter,

C_jc is the depletion capacitance of the collector-base junction per unit area,

N_D is the collector (donor) dopant concentration, assumed to be constant and much smaller than base dopant concentration N_A except in the immediate vicinity of the collector-base junction,

μ_n, which equals D_n q/kT, is the electron mobility,

q is the electronic charge,

k is Boltzmann's constant

T is the absolute temperature,

ε₀ is the permittivity of free space, and

K_S is the relative permittivity of silicon.

Use of t_BEFF in the dopant integral (in each) of Eqs. 2 and 4 indicates that this integral is taken across the quasi-neutral region of the intrinsic base--i.e., the region extending between the two depletion regions. The dopant integral in Eqs. 2 and 4 is commonly referred to as the Gummel number. The dopant integral in Eqs. 3 and 5 is taken across the full metallurgical thickness of the base as indicated by the use of t_B in this integral.

Eqs. 1-4 are available in prior art semiconductor literature. See: (a) Philips, Transistor Engineering (McGraw-Hill; reprinted: Robert E. Krieger Pub. Co., 1981), 1962, pages 298-304; (b) Warner et al, Transistor Fundamentals for the Integrated-Circuit Engineer (John Wiley & Sons), 1983, pages 559-562; (c) Muller et al, Device Electronics for Integrated circuits (John Wiley & Sons), 1977, pages 241-245; and (d) Grove, Physics and Technology of Semiconductor Devices (John Wiley & Sons), 1967, pages 228-230.

Eq. 5 has been derived here based on the following approximations: (a) the base region is totally depleted by the collector-base voltage at punch-through and (b) the collector-base junction is sufficiently asymmetrical to apply the single-sided abrupt approximation for the calculation of its depletion layer thickness, and to neglect the voltage drop on its heavily doped side. These modeling approximations are considered to be particularly valid in high-frequency transistors.

In particular, equating the depletion layer charges in the base and collector leads to: ##EQU2## where X_dC is the total depletion thickness on the collector side of the collector-base junction. Assuming that x_dC is approximately equal to the total depletion thickness along the collector-base junction according to classical pn-junction theory as provided in Grove (cited above), one has: ##EQU3## Substituting Eq. 7 into Eq. 6 produces Eq. 5.

If base dopant concentration N_A is reduced, the dopant integral in Eq. 2 decreases along with t_BEFF in Eq. 1. Collector current I_C and cutoff frequency f_T thereby increase, as is desirable. However, the dopant integrals in Eqs. 3-5 also decrease. This leads to decreases in punch-through voltage V_P and Early voltage V_A and an increase in base resistance r_B, all of which are disadvantageous. It would be desirable to increase the collector current and cutoff frequency without degrading the punch-through voltage, Early voltage, and base resistance.

GENERAL DISCLOSURE OF THE INVENTION

The present invention furnishes a special two-dimensional intrinsic base doping profile that enables the output current-voltage ("I-V") characteristics to be improved in a vertical bipolar transistor. For example, the magnitudes of the punch-through voltage and Early voltage are considerably higher in the transistor of the invention than in an otherwise similar bipolar transistor having the same active-region collector current extrapolated to zero collector-to-emitter voltage as the present transistor but lacking the special two-dimensional intrinsic base doping profile of the invention. Alternatively, the invention enables the collector current to be increased while maintaining the values of the punch-through and Early voltages.

When the magnitude of the punch-through voltage is raised at constant zero-V_CE extrapolated active-region collector current in accordance with the invention, the base resistance is typically reduced without decreasing the cutoff frequency. Alternatively, when the invention is employed to raise the collector current at constant punch-through voltage, the cutoff frequency typically increases without degrading the base resistance. In either case, the net result is a substantial improvement in transistor performance.

The intrinsic base of the present transistor includes a main intrinsic portion. The two-dimensional intrinsic base doping profile of the invention is achieved with a pair of more lightly doped base portions that encroach substantially into the intrinsic base below the main intrinsic base portion. The two deep encroaching base portions set up a two-dimensional charge-sharing mechanism. Part of the fixed collector depletion charge which would otherwise couple vertically with the fixed depletion charge in the main intrinsic portion of the base so as to cause punch-through now couples laterally with fixed depletion charge in the two encroaching base portions. For a given nominal base thickness, the magnitude of the punch-through voltage thereby increases.

More particularly, the bipolar transistor of the invention is furnished with an emitter, collector, and intervening multi-component base all situated in a vertical arrangement in a semiconductor body. The emitter extends to the upper surface of the semiconductor body. The main intrinsic portion of the base is located below the emitter and above a main portion of the collector.

As indicated above, the two base portions which encroach into the intrinsic base have lighter doping than the main intrinsic base portion and extend deeper into the semiconductor body than the main intrinsic base portion. The two deep encroaching base portions are laterally separated from each other below the main intrinsic base portion by a minimum spacing S_NE which is less than three times the minimum vertical thickness t_BMIN of the base. Preferably, S_NE is less than twice T_BMIN. In any case, the two encroaching base portions extend sufficiently close to each other to set up the above-mentioned charge sharing and thereby improve the output I-V characteristics.

Inasmuch as the two deep encroaching base portions are more lightly doped than the main intrinsic base portion, the upper boundary of the portion of the collector-base depletion region below the emitter extends to substantially the same vertical position in the intrinsic base of the present transistor as in the intrinsic base of a bipolar transistor that lacks the two deep encroaching base portions but is otherwise the same as the transistor of the invention. Thus, the presence of the two deep encroaching base portions does not significantly change the effective electrical thickness of the base. Accordingly, the cutoff frequency of the present transistor is substantially the same as that of an otherwise equivalent bipolar transistor that lacks the deep encroaching base portions.

The two deep encroaching base portions can be arranged in various configurations with respect to the emitter. When, as viewed in a direction perpendicular to the upper semiconductor surface, the emitter is in the shape of a stripe, the two nearest opposing edges of the deep encroaching base portions typically extend along the length of the stripe. Alternatively, the two nearest opposing edges of the deep encroaching base portions can cross the stripe, typically at a right angle. The transistor can also be furnished with one or more additional emitters and one or more additional encroaching base portions.

In fabricating the transistor of the invention, the following steps are performed on a surface-adjoining major region of a first conductivity type in a semiconductor body. Dopant of a second conductivity type opposite to the first conductivity type is introduced into the major region through its upper surface to define a first base layer of the second conductivity type. Dopant of the first conductivity type is introduced into the major region through a segment of its upper surface to define the emitter as a region of the first conductivity type overlying a main intrinsic portion of the first base layer. Further dopant of the second conductivity type is introduced into the major region through laterally separated segments of its upper surface. During this step, part of the further dopant is driven sideways and downwards to form the two deep encroaching base portions as regions of the second conductivity type.

The three dopant-introduction steps can be initiated in various orders. Typically, these three steps are performed at least partially in parallel.

In one embodiment of the present method, the step of introducing the further dopant to form the two deep encroaching base portions is initiated after initiating the other two dopant-introduction steps. In this case, the further dopant is typically implemented with a semiconductor impurity which diffuses through the major region at a faster rate than the dopants used in the other two dopant-introduction steps. When the first and second conductivity types respectively are n-type and p-type, aluminum is preferably utilized as the fast-diffusing impurity. Consequently, the locations of the encroaching base portions can be established without significantly affecting the locations of the emitter and the main intrinsic base portion. During the driving of the further dopant, spacers provided along an emitter contact are typically used to control the locations of the encroaching base portions.

Alternatively, the step of introducing the further dopant can be initiated before initiating the emitter dopant-introduction step.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a cross-sectional front view of a conventional vertical bipolar transistor.

FIGS. 2a and 3a are simplified one-dimensional diagrams of a bipolar transistor respectively in normal operation and in a punch-through condition.

FIGS. 2b and 3b are graphs for potential as a function of distance for the simplified transistor of FIGS. 2a and 3a respectively during normal operation and punch-through.

FIG. 4 is a graph of collector current as a function of collector-to-emitter voltage for indicating the Early voltage.

FIG. 5 is a cross-sectional front view of a vertical bipolar transistor that contains a pair of deep encroaching base portions in accordance with the invention.

FIG. 6 is a three-fold magnification of a portion of FIG. 5 generally centering around the main intrinsic base portion.

FIG. 7 is a layout view for the transistor of FIG. 5. The cross section of FIG. 5 is taken through plane 5-5 in FIG. 7.

FIG. 8 is a stepped cross-sectional side view for the transistor of FIG. 5. The cross section of FIG. 8 is taken through stepped plane 8--8 in FIGS. 5 and 7. The cross section of FIG. 5 is further taken through plane 5--5 in FIG. 8.

FIG. 9 is a graph of dopant concentration as a function of depth for an implementation of the transistor of FIG. 5.

FIG. 10 is a cross-sectional front view of a simplified vertical bipolar transistor provided with a pair of deep encroaching base portions in accordance with the invention for use in computer simulation.

FIGS. 11a and 11b are graphs for dopant concentration as a function of depth for the encroached-base transistor of FIG. 10.

FIG. 12 and 13 are three-dimensional graphs of the doping surfaces respectively for the encroached-base transistor of FIG. 10 and a simulated baseline transistor.

FIG. 14 is a graph for dopant concentration as a function of depth for two simulated transistors in a baseline simulation group. The two simulated transistors in FIG. 14 have the extreme dopant profiles considered in the baseline simulation group.

FIG. 15 is a graph for collector current as a function of collector-to-emitter voltage for all the simulated transistors in the baseline simulation group.

FIGS. 16a, 16b, and 16c are graphs illustrating lines of constant potential for different values of collector-to-emitter voltage in the encroached-base transistor of FIG. 10.

FIGS. 17a, 17b, and 17c are graphs for current flow at different values of collector-to-emitter voltage in the encroached-base transistor of FIG. 10.

FIGS. 18 and 19 are graphs for current flow, both electrons and holes, respectively for the encroached-base transistor of FIG. 10 and the simulated baseline transistor.

FIGS. 20 and 21 are graphs for the potential surfaces respectively for the encroached-base transistor of FIG. 10 and the simulated baseline transistor.

FIG. 22 is a graph for potential as a function of depth for the encroached-base transistor of FIG. 10 and the simulated baseline transistor.

FIGS. 23 and 24 are graphs for collector current as a function of collector-to-emitter voltage for the encroached-base transistor of FIG. 10 and the simulated baseline transistor for respective situations in which impact ionization is absent and present.

FIGS. 25 and 26 are three-dimensional graphs of the electric field surfaces respectively for the encroached-base transistor of FIG. 10 and the simulated baseline transistor.

FIG. 27 is a graph for collector current as a function of collector-to-emitter voltage for the encroached-base transistor of FIG. 10 and all the transistors in the baseline simulation group.

FIGS. 28a, 28b, 28c, 28d, 28e, and 28f are cross-sectional front views representing steps in a process for manufacturing the transistor of FIG. 5 in accordance with the invention.

FIG. 29 is a cross-sectional front view representing a step that can be substituted into the process of FIGS. 28a-28f for manufacturing a vertical bipolar transistor in accordance with the invention.

FIGS. 30, 32, 33, and 35 are partial transistor layout views for extensions of the layout of FIG. 7 in accordance with the invention.

FIG. 31 is a cross-sectional front view through plane 31--31 in FIG. 30.

FIG. 34 is a cross-sectional side view through plane 34--34 in FIGS. 33 and 35.

FIGS. 36, 37, 38, and 39 are partial mask layout views respectively corresponding to the partial transistor layout views of FIGS. 30, 32, 33, and 35.

Dashed lines generally indicate boundaries between regions of the same conductivity type but of substantially different dopant concentration. In the drawings containing depletion regions, dotted lines generally indicate the locations of the depletion regions. The metallurgical interface between substrate and epitaxial layer is indicated by dot-and-dash lines in the drawings.

Like reference symbols are employed in the drawings and in the description of the preferred embodiments to represent the same or very similar item or items.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

FIGS. 5-8 illustrate a narrow-emitter vertical npn transistor structure configured according to the teachings of the invention so as to improve the output I-V characteristics. In particular, the magnitudes of punch-through voltage V_P and Early voltage V_A are both greater in the transistor of FIGS. 5-8 than in an otherwise identical bipolar transistor that has either the same extrapolated active-region collector current at zero collector-to-emitter voltage or the same minimum base thickness but lacks the enhancements of the invention. Also, the base resistance r_B is typically reduced in the transistor of FIGS. 5-8 without impairing the transistor cutoff frequency f_T.

The cross-sectional views of FIGS. 5 and 8 are taken through planes 5--5 and 8--8 in the transistor layout view of FIG. 7. FIG. 6 is a magnification of the portion of FIG. 5 centered generally around the intrinsic base.

The transistor of FIGS. 5-8 has an emitter, collector, and intervening base. These elements are created from a semiconductor body consisting of a (100) lightly doped p-type monocrystalline silicon semiconductor substrate 50 and an overlying moderately doped n-type epitaxial silicon layer 52. P- substrate 50 has an average net dopant concentration of 2×10¹⁴ -6×10¹⁴ atoms/cm³, typically 4×10¹⁴ atoms/cm³. N epitaxial layer 52 has an average net dopant concentration of 1×10¹⁶ -4×10¹⁶ atoms/cm³, typically 2×10¹⁶ atoms/cm³. The thickness of epitaxial layer 52 is 0.9-1.2 m, typically 1.0 μm.

A patterned electrically insulating field region 54 of silicon oxide is recessed into the semiconductor body along the upper surface of n epitaxial layer 52. Field-oxide region 54 typically extends fully through epitaxial layer 52 and slightly into p- substrate 50 so as to divide layer 52 into a group of laterally separated active semiconductor device regions. Two such active device regions are shown in FIG. 5. The lateral dimension of the portion of field oxide 54 lying between these two device regions is not drawn to scale in FIG. 5.

A very heavily doped buried n-type layer 56 situated along the metallurgical interface between substrate 50 and layer 52 electrically interconnects the two active device regions shown in FIG. 5. Buried n++ layer 56 has a maximum net dopant concentration of 2×10¹⁹ -8×10¹⁹ atoms/cm³, typically 4×10¹⁹ atoms/cm³.

The transistor's emitter is a very heavily doped n-type zone 58 situated in the left-hand active region along the upper surface of the semiconductor body. N++ emitter 58 reaches a maximum net dopant concentration of 1×10²⁰ -4×10²⁰ atoms/cm³, typically 2.5×10²⁰ atoms/cm³,at the upper semiconductor surface Emitter 58 extends to a depth d_E of 0.09-0.11 μm, typically 0.10 μm, into epitaxial layer 52. Emitter depth d_E is also the emitter thickness (corresponding to emitter thickness t_E in the one-dimensional representation of FIG. 2a). The lateral width W_E of emitter 58 is 0.5-0.7 μm, typically 0.6 μm.

An overlying very heavily doped n-type polysilicon emitter contact 60 coated with a thin metal silicide cap 62 contacts emitter 58 in a self-aligned manner. Polysilicon emitter contact 60 has a lateral width W_ECONT of 0.45-0.55 μm, typically 0.5 μm.

The base of the transistor consists of a heavily doped p-type intrinsic base layer 64, a pair of moderately doped deep p-type encroaching base portions 66, and a pair of very heavily doped p-type extrinsic base contact zones 68 respectively corresponding to encroaching base portions 66. Components 64-68 are divided into an intrinsic base and a pair of extrinsic bases located symmetrically on opposite sides of the intrinsic base.

The intrinsic part of the base lies directly below n++ emitter 58 so that the lateral width W_IB of the intrinsic base is the same as emitter width W_E. Portions of

components

64 and 66 form the intrinsic base. P+ base layer 64 extends fully across the left-hand active semiconductor region in FIG. 5 and reaches a maximum net dopant concentration of 2×10¹⁸ -8×10¹⁸ atoms/cm³, typically 5×10¹⁸ atoms/cm³, at the upper semiconductor surface. Layer 64 extends to a depth d_IB of 0.18-0.24 μm, typically 0.20 μm, into n epitaxial layer 52.

P base portions 66 extend deeper into n epitaxial layer 52 than p+ base layer 64 and have a lighter doping concentration than p layer 64. In particular, p portions 66 reach a maximum net dopant concentration of 1×10¹⁷ -4×10¹⁷ atoms/cm³, typically, 2×10¹⁷ atoms/cm³, along the bottom of layer 64 outside the intrinsic base. Each portion 66 extends to a depth d_DB of 0.4-0.6 μm, typically 0.45 μm, into epitaxial layer 12 outside the intrinsic base. FIG. 9 illustrates a typical doping profile for the transistor of FIG. 5 except for p++ base contact zones 68.

A main portion 64M of p+ base layer 64 directly underlies n++ emitter 58 and constitutes part of the intrinsic base. Each deep p portion 66 encroaches substantially into the semiconductor material directly below p+ main intrinsic base portion 64M to form an additional part of the intrinsic base. The minimum thickness of the intrinsic base occurs at a location between encroaching base portions 66 and is also the minimum thickness t_BMIN of main intrinsic base portion 64M. Inasmuch as t_BMIN equals d_IB -d_E, t_BMIN equals 0.08-0.12 μm, typically 0.10 μm.

Deep encroaching base portions 66 extend sufficiently close to each other below main intrinsic base portion 64M, without touching one another along their nearest side edges 66E (see FIG. 7), that the electric field in the transistor structure is shaped in such a way as to increase the magnitude of punch-through voltage V_P beyond what would exist if encroaching portions 66 were absent. An increase of more than 20% in the magnitude of voltage V_P is typically achievable in the encroached-base transistor of FIG. 5. This is a clear improvement.

To achieve such punch-through voltage improvement, the minimum spacing S_NE between encroaching base portions 66 is less than three times, usually less than two times, T_BMIN. Preferably, S_NE is less than 1.5t_BMIN. S_NE is typically in the vicinity of 0.1 μm when emitter width W_E and the base doping have values in the vicinity of those given above for the transistor of FIG. 5. The precise value of spacing S_NE is critically determined by simulation and experimentation so as to obtain a desired increase in the V_P magnitude.

Each extrinsic base consists of one of p++ base contact zones 68, the adjoining part of p+ intrinsic base layer 64 outside main intrinsic base portion 64M, and the underlying part of p encroaching base portion 66. In addition to performing the field-shaping function that improves the V_P characteristics, encroaching base portions 66 serve as connection zones that help provide low-resistance paths from the intrinsic base to base contact zones 68 in the extrinsic bases.

Contact zones 68 reach a maximum net dopant concentration of 1×10¹⁹ -5×10¹⁹ atoms/cm³, typically 2×10¹⁹ atoms/cm³, at the upper semiconductor surface. Although FIG. 5 indicates that contact zones 68 do not extend as deep into n epitaxial layer 52 as intrinsic base layer 64, zones 68 could extend deeper into epitaxial layer 52 than layer 64. However, encroachment of heavily doped contact zones 68 into the intrinsic base needs to be avoided.

Each base contact zone 68 contacts an overlying very heavily doped p-type polysilicon base contact 70. In the illustrated example, parts of p+ base layer 64 situated in the extrinsic bases also make contact with base contact 70. A thin metal silicide cap 72 extends from contact zone 68 to base contact 70 to reduce the base resistance. As shown in FIGS. 7 and 8, base contact 70 and cap 72 are in the shape of a "U" that extends from one of the extrinsic bases to the other.

A pair of electrically insulating spacers 74 consisting of silicon oxide are situated along the sidewalls of emitter contact 60. Oxide spacers 74 laterally separate metal silicide cap 72 from emitter contact 60 and metal silicide cap 62 so as to avoid short circuiting the base to emitter 58. During transistor fabrication, spacers 74 are typically utilized to control the lateral separation between emitter 58 and each base contact zone 68. As described below, spacers 74 are also employed to control spacing S_NE during the formation of encroaching base portions 66. Each spacer 74 typically has a bottom width W_S (taken in the same direction as emitter width W_E) of 0.1-0.2 μm, typically 0.15 μm.

The transistor's collector includes a main collector region formed with n++ buried layer 56 and an overlying moderately doped n-type epitaxial portion 76 that typically extends up to main intrinsic base portion 64M and encroaching base portions 66 in the left-hand active semiconductor region in FIG. 5. The collector also includes a very heavily doped n-type collector contact region 78 that provides the main collector region with an electrical path to the upper semiconductor surface through the right-hand active semiconductor region in FIG. 5. An overlying very heavily doped n-type polysilicon collector contact 80 coated with a thin metal silicide cap 82 contacts n++ collector contact 78 to complete the transistor structure.

The improvement in punch-through voltage V_P occurs for the following reasons. Punch-through in a vertical bipolar transistor having symmetrically located extrinsic bases normally occurs along the center plane of the intrinsic base where the depletion region of the collector-base junction first touches the depletion region of the emitter-base junction when collector-to-emitter voltage V_CE is high enough. In the transistor of FIG. 5 where the emitter is narrow and the intrinsic base is substantially encroached by deep p base portions 66, a portion of the fixed positive depletion charge in the center of the collector part of the depletion region along the collector-base junction is shared with (a) fixed negative depletion charge in the main intrinsic part of the depletion region along the collector-base junction (vertical charge coupling) and with (b) fixed negative depletion charge in the encroaching base part of the depletion region along the collector-base junction (lateral charge coupling). FIG. 6 illustrates the charge sharing.

Due to the charge sharing, only part of the collector depletion charge enters the charge balance relationship that determine punch-through voltage V_P. Accordingly, Eq. 6 above is changed to: ##EQU4## where α, which is less than 1, is a charge-sharing coefficient. Eq. 5 therefore changes to: ##EQU5##

Inasmuch as α is less than 1, the strong encroachment of p regions 66 into the intrinsic base increases voltage V_P over the value that would otherwise occur if p regions 66 were absent (i.e., did not encroach). In physical terms, it takes more voltage on the collector-base junction to create, by depletion, the positive charge necessary to balance the negative charge corresponding to the punch-through condition in the central portion of the fully depleted base.

The distance from the bottom of emitter 58 to the underlying point along the collector-base junction is the (vertical) base thickness t_B. As shown in FIG. 6, the presence of deep encroaching base portion 66 causes metallurgical base thickness t_B to vary greatly across the intrinsic base.

The dotted lines in FIG. 6 indicate the approximate boundaries of the depletion regions of the collector-base and emitter-base junctions. In particular, dotted line 84 represents the approximate boundary of the base portion of the depletion region of the collector-base junction. Dotted line 86 similarly represents the approximate boundary of the base portion of the depletion region of the emitter-base junction. The region between

depletion boundaries

84 and 86 is thus the quasi-neutral portion of the intrinsic base. The distance between

boundaries

84 and 86 is the effective thickness t_BEFF of the quasi-neutral base region.

Inasmuch as deep encroaching base portions 66 are more lightly doped than base layer 64, depletion boundary 84 is situated in base layer 64 relatively close to the bottom of layer 64. Accordingly, effective base thickness t_BEFF is substantially constant across the intrinsic base even though actual base thickness t_B varies greatly across the intrinsic base. Importantly, the portion of depletion boundary 84 located below emitter 58 extends to substantially the same vertical position that would arise if encroaching base portions 66 were absent. With reference to Eq. 1, cutoff frequency f_T is thus of substantially the same value that would arise in a vertical bipolar transistor lacking encroaching base portions 66 but otherwise identical to the encroached-base transistor of FIGS. 5-8.

Instead of utilizing the charge-sharing mechanism to increase punch-through voltage V_P, the dimensional/doping parameters of the transistor can be modified to reduce minimum base width T_BMIN. Collector current I_C and cutoff frequency f_T can then be increased while voltage V_P is maintained constant. Alternatively, the transistor's dimensional/doping parameters can be adjusted to provide intermediate improvements in voltage V_P and current I_C. Like comments apply to the Early voltage V_A. In these ways, the invention improves the output I-V characteristics of a vertical bipolar transistor having an intrinsic base situated between a pair of extrinsic bases.

Computer simulations have been done to analyze the effect of the charge sharing in a vertical bipolar transistor having its intrinsic base substantially encroached in the manner described above. The simulations were performed with the Medici two-dimensional semiconductor device simulation program available from Technology Modeling Associates. The simulated transistors were described on a dense 3500-node grid using half-device representation with reflective boundary conditions on both sides of the representation.

FIG. 10 illustrates the profile of an encroached-base npn transistor structure employed in the computer simulations. The left-hand axis in FIG. 10 extends along the centerline of the emitter and base. To simplify the analysis, the p++ base contact zones were omitted in the simulations, and the base contacts were forced to be ohmic. The collector contact was assumed to be located at the bottom of the structure along the emitter-base centerline. Also, the upper semiconductor surface was assumed to be substantially flat.

Items

90 and 92 in FIG. 10 respectively indicate the polysilicon base contact and a simplified version of spacer 74.

For the computer simulations, it was assumed that the dopings for emitter 58 and intrinsic base layer 64 were initiated before the doping for deep encroaching base portions 66. It was also assumed that the two-dimensional dopant concentration profile of encroaching base portions 66 resulted from diffusing a fast-diffusing p-type dopant, such as aluminum, into the semiconductor body from a dopant source situated next to spacer 92. Accordingly, deep encroaching base portion 66 in FIG. 10 was self-aligned to the outside edge of spacer 92, and was formed without significantly affecting the dopant profiles for shallow emitter 58 and intrinsic base layer 64.

The depth of lateral diffusion was assumed to be approximately 80% of the depth of vertical diffusion for both n-type and p-type dopants. Spacing S_NE was thus given approximately as:

S.sub.NE =W.sub.E +2W.sub.S -2(0.8d.sub.DB)                (10)

In a similar manner, polysilicon emitter contact 60 was assumed to constitute a dopant source for emitter 58 so that emitter contact 60 was self-aligned to the emitter 58.

The following numerical values were used in the simulations:

n-type emitter dopant concentration at upper surface=2.5×10²⁰ atoms/cm³,

p-type dopant concentration of intrinsic base layer 64 at upper surface =5.0×10¹⁸ atoms/cm³,

upper-surface concentration of p-type dopant that produces encroaching base portions 66=2.0×10¹⁷ atoms/cm³,

n-type dopant concentration of epitaxial layer 76 and lightly doped portion of collector=2.0×10¹⁶ atoms/cm³,

emitter depth d_E =0.10 μm,

depth d_IB of intrinsic base layer 64=0.20 μm, whereby minimum vertical base thickness t_BMIN =0.10 μm,

depth d_DB of encroaching base portions 66=0.50 μm,

epitaxial thickness=1.0 μm with standard gradual transition to n++ buried layer,

width W_ECONT of emitter contact 60=0.50 μm,

emitter width W_E =intrinsic base portion width W_IB =0.66 μm (as a result of 0.8 lateral-to-vertical diffusion ratio),

spacer width W_S =0.16 μm,

width W_BCONT of base contact 90=0.39 μm,

lateral width of base (total)=1.60 μm,

recombination-generation lifetime at low doping concentration in monocrystalline silicon for both electrons and holes=0.1 μsec,

recombination-generation lifetime at low doping concentration in polysilicon for both electrons and holes=0.01 μsec,

modeling parameter N_SRHN in the concentration-dependent recombination law=4×10¹⁷ atoms/cm³,

Auger recombination parameters AUGN and AUGP=2×10^-31 recombinations/cm³ -sec,

band gap narrowing parameter NO.BGN=1×10¹⁸ atoms/cm³, and

contact resistance at base silicide contact=40 ohm-μm².

All other material properties were set at their default Medici values.

FIG. 11a shows the dopant concentration of the encroached-base transistor simulated in FIG. 10 along the device centerline (X=0) through the intrinsic base. FIG. 11b shows the dopant concentration of the encroached-base transistor along a vertical line extending through the middle of the extrinsic base (X=W_ECONT /2+W_S +W_BCONT /2). FIG. 12 illustrates a three-dimensional representation of the dopant concentration in the encroached-base transistor of FIG. 10. Arrow BE in FIG. 12 indicates the dopant-concentration minimum that represents the location of the emitter-base junction.

The simulations on the transistor of FIG. 10 were compared to Medici simulations performed on a baseline vertical bipolar transistor configured in the same way as the simulated encroached-base transistor except that encroaching base portions 66 were deleted. The values of all the other parameters in the simulations on the baseline transistor were the same as those in the simulations on the encroached-base transistor. FIG. 13 presents a three-dimensional representation of the dopant concentration in the baseline transistor. Arrow BE again indicates the dopant-concentration minimum representing the emitter-base junction.

The depth of the intrinsic base layer in the baseline transistor was 0.20 μm, just as in the encroached-base simulated transistor. Because the baseline transistor lacked deep encroaching base portions 66, the 0.20-μm depth of the intrinsic base layer was also the depth of the substantially flat collector-base junction in the baseline transistor. Accordingly, base thickness t_B of the baseline transistor was also its minimum vertical base thickness. Since emitter thickness t_E was 0.10 μm, both t_BMIN for the encroached-base transistor of FIG. 10 and (minimum) base thickness t_B for the baseline transistor equaled 0.10 μm in the simulations.

For additional comparison purposes, a computer simulation experiment was performed in which the doping profile of a vertical bipolar transistor having a substantially flat collector-base junction situated at a depth of 0.25 μm but lacking the deep encroaching base portions of the present invention was adjusted in such a way as to reduce the collector-base junction depth in five 0.01-μm steps, thereby reaching the baseline transistor. These five steps correspond to reducing base thickness t_B from approximately 0.15 μm to the 0.10-μm value of our baseline transistor. The values of all the other parameters were the same as in the simulations on the encroached-base transistor. The six simulated transistors which lacked the encroached-base portions of the invention are referred to here collectively as the "baseline simulation group". FIG. 14 depicts the centerline dopant profile for the particular baseline simulation-group transistors having the 0.15-μm and 0.10-μm base thicknesses--i.e., the extreme values of the t_B range considered in the simulations.

The results of the simulations on the baseline simulation group are summarized in FIG. 15 in terms of the output I-V characteristics--i.e., collector current I_C as a function of collector-to-emitter voltage V_GE --for all six transistors in the baseline simulation group. Base-emitter voltage V_BE was fixed at 0.8 V. In light of the consideration/non-consideration of impact ionization in certain of the simulations as discussed further below, note that the impact ionization of the Medici program was employed in the simulations of FIG. 15.

Reducing base thickness t_B from 0.15 μm to 0.10 μm in the baseline simulation group caused collector current I_C to increase by a factor of approximately 2.5 as indicated in FIG. 15. However, the slopes of the output I-V curves increase greatly with decreasing t_B. This is undesirable.

For example, at a V_CE of 5 V, the slope of the I-V curve for the baseline transistor (0.10-μm t_B) was over 15 times the slope of the I-V curve for the transistor having the 0.15-μm t_B. Accordingly, the baseline transistor had much lower Early voltage V_A and punch-through voltage V_P than the transistor whose t_B was 0.15 μm. In short, reducing base thickness t_B in the baseline simulation group in an effort to increase collector current I_C and cutoff frequency f_T caused voltages V_P and V_A to be substantially degraded.

FIGS. 16a-27 illustrate the simulated performance of the encroached-base transistor of FIG. 10, partly in comparison with the baseline transistor. Unless otherwise indicated, base-emitter voltage V_BE is 0.8 V in each of FIGS. 16a-27. In order to differentiate between the simultaneous effects of impact ionization generation and direct current injection from the emitter-base junction, the simulations in FIGS. 16a-23 were done with the impact ionization turned off in the Medici software. Impact ionization was employed in the simulations of FIGS. 24-27.

The electrostatic operation of the encroached-base transistor of FIG. 10 is illustrated in FIGS. 16a-16c (collectively "FIG. 16") for collector-to-emitter voltage V_CE respectively at 1, 5, and 12 V. The solid lines in FIG. 16 represent the lines of constant potential at the indicated steps (ΔV) in potential. From an examination of FIG. 16, it is apparent that the thickness t_BEFF of the electrically neutral region of the intrinsic base does not vary significantly, at least up to V_CE equal to 5 V, despite the large variation in the actual thickness t_B of the metallurgical intrinsic base.

Furthermore, the simulations show that the average electrical thickness of the intrinsic base in the encroached-based simulated transistor is 15-25% less than the average electrical thickness of the intrinsic base in the baseline transistor. Using the transistor of the present invention, properties such as cutoff frequency f_T, collector current I_C, and the current gain, which require a thin base are at least as good, and possibly better, than in the baseline transistor even though the metallurgical base in the invention is significantly thicker than in the baseline transistor.

FIG. 17a-17c (collectively "FIG. 17") illustrate the electrokinetic operation of the encroached-base transistor of FIG. 10 at the same respective values of collector-to-emitter voltage V_CE as in FIG. 16. The solid lines in FIG. 17 represent current-flow contours at 5% increments (ΔI). As FIG. 17 shows, current moves in the vertical direction with little non-uniformity at the injecting emitter-base junction even though there is a large non-uniformity in metallurgical base thickness t_B across the intrinsic base.

FIG. 18 illustrates an extension of FIG. 17b to which vector plots of hole current are added for the encroached-base transistor of FIG. 10. FIG. 19 presents a composite representation of the vector hole current for the baseline transistor at the same V_CE value as in FIG. 18. The current spreading in the encroached-base transistor is greater than in the baseline transistor. This indicates that encroaching base portions 66 contribute to lowering the base resistance r_B of the encroached-base transistor.

FIGS. 20 and 21 present three-dimensional representations of the potential that electrons see in moving respectively through a central region of the encroached-base transistor of FIG. 10 and a corresponding central region of the baseline transistor. Collector-to-emitter V_CE is 12 V in both of FIGS. 20 and 21.

The charge-sharing effect responsible for the improved performance in the present invention can be seen by comparing FIG. 20 to FIG. 21. As indicated by arrow PD in FIG. 20, the potential surface of the encroached-base transistor is pushed down near the center of the emitter-base junction, thereby flattening the bottom of the potential surface and raising the magnitude of the potential barrier that electrons must overcome to enter the base. On the other hand, FIG. 21 shows that the baseline transistor has a non-distorted potential distribution with a less-flattened bottom and a smaller barrier potential. The differences in shape of the potential surface and in magnitude of the barrier potential to electron flow can be seen even more clearly in FIG. 22 which depicts the potential distributions along the centerlines of both transistors.

The relatively flat bottom of the potential distribution shown in FIGS. 20 and 22 for the encroached-base transistor corresponds to diffusion-limited resistance to electron flow through the base. Because diffusion limits electron movement in the base, collector-to-emitter voltage V_CE has only a small influence on the magnitude of collector current I_C. That is, the flattened bottom of the potential curve for the encroached-base transistor represents the quasi-neutral base region which, as can be seen by comparison to FIG. 2b, occurs when the transistor is not in a punch-through condition.

In contrast to the flattened bottom of the potential surface in the encroached-base transistor, FIGS. 21 and 22 show that the potential distribution for the baseline transistor rises rapidly after passing a potential minimum. The less-flattened bottom of the potential distribution indicates that there is less diffusion resistance to electron movement through the base. Collector-to-emitter voltage V_CE strongly controls base-to-emitter voltage V_BE and readily pulls the potential curve upward to reduce the magnitude of the barrier potential. The baseline transistor is much closer to punch-through than the encroached-base transistor of FIG. 10.

Along the centerlines of the two transistors, the potential for the encroached-base transistor drops 20-30 mvolts lower than the potential for the baseline transistor as shown in FIG. 22. While this is only about 10% of the total potential barrier at the emitter-base junctions of the two transistors, the effect on the current flow is large because the current varies exponentially with the magnitude of the barrier potential. Moreover, collector-to-emitter voltage V_CE is directly applied to the potential barrier in the baseline transistor, thereby causing its collector current I_C to rise rapidly with increasing V_CE. FIGS. 20-22 thus show that the magnitude of punch-through voltage V_P is greater in the encroached-base transistor than in the baseline transistor. The same applies to Early voltage V_A.

FIG. 23 directly compares the output I-V characteristics of the encroached-base transistor of FIG. 10 and the baseline transistor. Although collector current I_C is somewhat higher in the baseline transistor, the slope of the I-V curve for the encroached-base transistor is much less than for the baseline transistor. This likewise shows that the encroached-base transistor has much better punch-through voltage V_P and Early voltage V_A than the baseline transistor. Since the impact ionization was turned off for the simulation in FIG. 23, it is clear that the increased slope of the simulated I-V characteristics in the baseline transistor is associated with the baseline transistor approaching a punch-through condition.

Impact ionization in the base of a bipolar transistor produces avalanche multiplication of charge carriers when the V_CE magnitude is raised to a sufficiently high magnitude. As with punch-through, avalanche charge-carrier multiplication causes the number of charge carriers that move across the collector-base junction to increase rapidly with increasing V_CE magnitude. The base essentially breaks down. The collector-to-emitter breakdown voltage is the V_CE value at which collector current IC increases rapidly with increasing V_CE magnitude due to a mechanism such as punch-through or avalanche multiplication.

The I-V simulation of FIG. 23 was, as mentioned above, performed with the impact ionization turned off in the Medici program. FIG. 24 presents a corresponding I-V simulation with the impact ionization turned on. Comparison of FIG. 24 to FIG. 23 indicates that the output I-V characteristics of the encroached-base transistor relative to those of the baseline transistor are even better when the impact ionization is employed in the simulations. The I-V curve for the encroached-base transistor is not changed significantly, whereas the I-V curve for the baseline transistor rises more rapidly when the impact ionization is present. This indicates that the magnitude of the collector-to-emitter breakdown voltage in the baseline transistor is close to or less than the magnitude of its punch-through voltage V_P when impact ionization is taken into consideration,

The apparent reason for the preceding differences is that, compared to the baseline transistor, the electric field for the encroached-based transistor is shifted towards lower values, thereby making it considerably less sensitive to impact ionization over the range of applied voltages of interest. This is confirmed by the simulations of FIG. 25 and 26 which respectively depict three-dimensional representations of the electric fields in the two simulated transistors at a V_CE of 12 V. The dopant distribution around the collector-base junction is more graded in the encroached-based transistor than in the baseline transistor. The less abrupt field variation in the encroached-base transistor is propagated into its intrinsic base.

FIG. 27 presents output I-V characteristics for the encroached-based transistor and all six transistors in the baseline simulation group for the case in which the impact ionization is present. In addition to having better output I-V characteristics than the baseline transistor, the encroached-based transistor has better output I-V characteristics than any of the other five transistors in the baseline simulation group.

For example, across the 0-to-12-V V_CE range of interest, the encroached-based transistor of FIG. 10 has approximately the same I-V slope as the 0.15-μm t_B baseline simulation-group transistor but provides approximately twice the collector current. At the other end of the t_B spectrum considered in the simulations, the extrapolated active-region value of collector current I_C at zero V_CE is only slightly less (1.0×10^-5 A/μm as indicated by arrow A in FIG. 27 compared to 1.3×10^-5 A/μm as indicated by arrow B in FIG. 27) in the encroached-base transistor of FIG. 10 than in the baseline transistor whose (minimum) base thickness t_B equals minimum base thickness t_BMIN of the encroached-base transistor. However, the slope of the I-V curve is much greater (a factor of 5-10 greater) for the baseline transistor than for the encroached-base transistor. Importantly, FIG. 27 indicates that the magnitude of the collector-to-emitter breakdown voltage for the baseline transistor is 10-11 volts. This is too low for many applications where the power supply voltage is 5 volts.

Comparison of the encroached-base transistor of FIG. 10 to the simulation-group transistor having the 0.11-μm t_B is quite useful. Although there is a 0.01 μm difference in the minimum thicknesses of the bases of these two transistors, their zero-V_CE extrapolated active-region collector currents are the same (1.0×10^-5 A/μm as indicated by arrow A in FIG. 27). But, the slope of the I-V curve for the 0.11-μm t_B simulation-group transistor is approximately three times the slope of the I-V curve for the encroached-base transistor. As a result, the collector-to-emitter breakdown voltage is over 20% higher in the encroached-base transistor. This clearly demonstrates the improvement attainable with the invention. The encroached-base transistor furnishes I-V output characteristics which, as a group, are better than in any of the six transistors in the baseline simulation group.

FIGS. 28a-28f (collectively "FIG. 28") illustrate a process for manufacturing the vertical bipolar transistor of FIG. 5 in accordance with the invention. The process of FIG. 28 is typically a modification of a BiCMOS fabrication process such as that described in Ilderem et al, U.S. Pat. No. 5,079,182. Accordingly, n-channel and p-channel insulated-gate field-effect transistors are typically formed as part of an integrated circuit that contains the bipolar transistor of FIG. 5. For simplicity, FIG. 28 and the accompanying process description do not deal with the fabrication of any transistors except that of FIG. 5.

The n-type dopings in the fabrication process are performed with antimony, arsenic, and/or phosphorus. The p-type dopings are done with boron and aluminum.

The starting point is p- substrate 50. Using a suitable masking technique, an n-type dopant is introduced into substrate 50 at the location for n++ buried layer 56. N epitaxial layer 52 is grown on substrate 50. During the epitaxial growth, part of the n-type dopant introduced into substrate 50 outdiffuses to produce buried layer 56 along the interface between substrate 50 and epitaxial layer 52 as shown in FIG. 28a.

Field-oxide region 54 is formed along the upper semiconductor surface according to a conventional technique. Using a photoresist mask (not shown), an n-type dopant is introduced into the right-hand active semiconductor region to convert its n epitaxial material into n++ collector contact zone 78. FIG. 28b shows the structure at this stage.

A layer 102 of polysilicon is deposited on top of the structure. A thin cap layer 104 of silicon oxide is formed along the top of polysilicon layer 102 by thermal oxidation. A p-type boron-containing dopant is ion implanted through cap layer 104 into polysilicon layer 102 at a moderate dosage. The structure is subsequently annealed at a temperature, and for a time, sufficient to cause some of the just-introduced boron to diffuse out of polysilicon layer 102 into epitaxial layer 52 to form a precursor to intrinsic base layer 64 in the left-hand active semiconductor region as indicated in FIG. 28c.

Using a photoresist mask (not shown), an additional p-type boron-containing dopant is ion implanted at a very heavy dosage through cap layer 104 into polysilicon layer 102 at locations for base contact 70. Similarly, an n-type dopant is ion implanted at a very heavy dosage through cap layer 104 into polysilicon layer 102 at locations for emitter contact 60 and collector contact 80 using a suitable photoresist mask (not shown). Using another photoresist mask (not shown), selected parts of

layers

104 and 102 are removed to define

contacts

60, 70, and 80 in the remaining polysilicon as shown in FIG. 28d.

Items

104A, 104B, and 104C respectively indicate the remaining portions of cap layer 104

overlying polysilicon contacts

60, 70, and 80.

Another anneal is typically performed. This anneal may be a separate bipolar step or part of a CMOS process. In any case, part of the n-type dopant in polysilicon emitter contact 60 outdiffuses to form a precursor to emitter 58. Some of the boron in polysilicon base contact 70 may also outdiffuse into p+ layer 64. Such p-type diffusion is not particularly material here and thus is not illustrated in the drawings. FIG. 28d illustrates the structure at the end of this anneal.

A layer of low-temperature oxide is deposited on the upper surface of the structure and anisotropically etched to form spacers 74 along the sidewalls of polysilicon emitter contact 60. Oxide spacers that form simultaneously along the sidewalls of

polysilicon contacts

70 and 80 are normally removed using a masked etching procedure. During the etch, a small thickness of the monocrystalline silicon at the exposed areas is typically removed.

Precursors to base contact zones 68 are formed by a procedure in which a p-type boron-containing dopant is introduced into epitaxial layer 52 through exposed parts of its upper surface using spacers 74 and polysilicon emitter contact 60 as a mask to control the lateral growth of p++ base contact zones 68 towards each other. This dopant introduction step can be performed by ion implantation or diffusion.

The structure is subsequently annealed to cause part of the just-introduced boron to diffuse sideways and downward to form the precursors to base contact zones 68. During the anneal, emitter 58 and intrinsic base layer 64 also expand, but to a lesser extent than contact zones 68. FIG. 29e shows the resulting structure.

Deep encroaching base portions 66 are subsequently formed by a procedure in which a fast-diffusing p-type dopant, such as aluminum, is ion implanted into epitaxial layer 52 through exposed parts of its upper surface using spacers 74 and polysilicon emitter contact 60 as a mask to control the lateral growth of encroaching base portions 66 towards each other. Aluminum is utilized as the p-type dopant for creating encroaching base portions 66 because the diffusion coefficient of aluminum is approximately six times the diffusion coefficient of boron.

The structure is then annealed to form deep encroaching base portions 66 from the just-introduced aluminum. During the anneal,

components

58, 64, 66, and 68 move to their final positions as shown in FIG. 28f. Since the aluminum utilized to form encroaching base portions 66 diffuses much faster than the boron employed to create p+ layer 64 and p++ base contact zones 68, the boron diffusion to establish the final positions for

components

64 and 68 is much less than the aluminum diffusion to form encroaching base portions 66. In essence, the process is designed in such a way that the formation of

components

58, 64, and 68 is largely independent of the formation of encroaching base portions 66.

Encroaching base portions 66 and base contact zones 68 are self-aligned to spacers 74 and polysilicon emitter contact 60 in the process of FIG. 28. In addition, n++ emitter contact 60 is self-aligned to emitter 58. As a result,

components

66 and 68 are self-aligned to emitter 58.

Metal silicide caps 62, 72, and 82 are provided in a conventional way by depositing a layer of a suitable metal on the upper surface of the structure and heating the structure to convert part of the metal into

caps

62, 72, and 82. The unreacted metal is removed to produce the structure of FIG. 5.

The process of FIG. 28 can be modified in various ways. For example, the p-type dopant for encroaching base regions 66 could be introduced into epitaxial layer 52 prior to introducing the dopant for base contact zones 68. The last two anneals could be combined into one anneal performed after the dopants for

components

68 and 66 are introduced into the substrate. In any case, the value of spacing S_NE between encroaching base portions 66 depends on the location of the outside bottom edges of spacers 74. The magnitude of spacing S_NE is thereby controlled by appropriately adjusting W_E, W_S, and/or d_DB.

FIG. 29 illustrates how the process of FIG. 28 can be modified to enable a special mask 110 to be employed in controlling the locations of encroaching base portions 66 instead of using spacers 74 and emitter contact 60 for this purpose. In applying the variation of FIG. 29 to the process of FIG. 28, the process steps are the same up to the stage shown in FIG. 28b. At this point, mask 110 is formed on top of the structure as illustrated in FIG. 29. Mask 110 may consist of photoresist.

A p-type boron-containing dopant is introduced into epitaxial layer 52 through the exposed parts of the upper semiconductor surface at locations intended for encroaching base portions 66. Mask 110 thus controls the lateral growth of base portions 66 towards each other. An anneal is subsequently performed to drive the boron sideways and downwards.

The structure of FIG. 29 can now be processed in the way described above for FIGS. 28c-28e. The resultant structure appears very much like the structure of FIG. 28f, except that encroaching base portions 66 are no longer self-aligned to emitter 58. Use of the process modification involving FIG. 29 alleviates the need to use a fast-diffusing dopant, such as aluminum, to create deep encroaching base portion 66. Instead, encroaching base portions 66 are formed with conventional boron prior to the creation of gate layer 58, emitter 64, and base contact zones 68. Thus, the comparatively greater annealing time required for creating deep encroaching base portions 66 does not advance the shallower dopant profiles of

components

58, 64, and 68.

Emitter 58 in the transistor of FIGS. 5-8 is in the shape of a stripe having length L_E greater than width W_E as viewed in a direction perpendicular to the upper semiconductor surface. The transistor of FIGS. 5-8 can be expanded to include two or more such emitter stripes.

FIG. 30 illustrates the layout for such an extension in which the so-extended transistor contains an additional very heavily doped n-type emitter stripe 58A that runs parallel to emitter stripe 58. FIG. 31 depicts a cross section for this transistor. The extended transistor of FIGS. 30 and 31 also contains an additional heavily doped p-type main intrinsic base portion 64MA, an additional moderately doped p-type encroaching base portion 66A, and one or two additional heavily doped p-type base contact zones 68A.

In addition to the intrinsic base formed with main intrinsic base portion 64M and the underlying parts of encroaching base portion 66, the transistor in FIGS. 30 and 31 has an intrinsic base formed with additional p+ main intrinsic base portion 64MA and the underlying parts of additional p portion 66A and center p portion 66. The minimum thickness t_BMINA of the additional intrinsic base occurs at a location between additional portion 66A and center portion 66.

Additional portion 66A and center portion 66 encroach substantially into the additional intrinsic base. In particular, additional portion 66A and center portion 66 extend sufficiently close to each other below additional main intrinsic base portion 64MA without touching along nearest side edges 66EA that the output I-V characteristics associated with additional intrinsic base portion 64MA, additional encroaching base portion 66A, and center encroaching base portion 66 largely match the improved output I-V characteristics attributable to intrinsic base portion 64M and the two encroaching base portions 66.

To achieve this match, minimum thickness t_BMINA of the additional intrinsic base largely equals T_BMIN. Also, the minimum spacing S_NEA between additional encroaching base portion 66A and center encroaching base portion 66 largely equals S_NE. Accordingly ratio S_NEA /t_BMINA largely equals ratio S_NE /t_BMIN . Hence, spacing S_NEA is less three times, usually less than twice, t_BMINA . Preferably, S_NEA is less than 1.5t_BMINA. Two-dimensional simulations and experimentation are utilized to set the values of spacings S_NE and S_NEA according to the desired amount of output I-V improvement.

In the layouts of FIGS. 7 and 30, nearest side edges 66E of encroaching base portions 66 extend parallel to emitter stripe 58. Alternatively, edges 66E could cross emitter stripe 58.

FIG. 32 illustrates a modification of the transistor layout of FIG. 7 in which side edges 66E cross emitter stripe 58 at a right angle. Although part of the geometry is rotated 90°, the S_NE /t_BMIN relationship presented above for the layout of FIG. 7 applies directly to the layout of FIG. 32.

The transistor layout of FIG. 32 can be expanded to include one or more additional encroaching base portions. FIG. 33 illustrates such an extension where there is one additional encroaching base portion 66B. FIG. 34 depicts a cross section for this extension. The nearest side edges 66EB of additional portion 66B and center portion 66 cross emitter stripe 58 at a right angle. Unlike the transistor of FIGS. 30 and 31 where there is an additional intrinsic base, the transistor of FIGS. 33 and 34 has only a single intrinsic base and thus only main intrinsic base portion 64M.

Additional portion 66B and center portion 66 in FIGS. 33 and 34 encroach substantially into the intrinsic base. They extend sufficiently close to each other below main intrinsic base portion 64M without touching one another along nearest side edges 66EB that the output I-V characteristics associated with additional encroaching base portion 66B and center encroaching base portion largely match the improved output I-V characteristics attributable to the two encroaching base portions 66.

The preceding match is attained by arranging for the minimum spacing S_NEB between additional encroaching base portion 66B and center encroaching base portion 66 to largely equal S_NE. Hence, ratio S_NEB /t_BMIN largely equals ratio S_NE /t_MIN. Accordingly, spacing S_NEB is less than three times, usually less than twice, t_BMIN. S_NEB preferably is less than 1.5t_BMIN.

FIG. 35 illustrates an expansion of the layout of FIG. 33 to include additional emitter stripe 58A. The cross section of FIG. 34 applies to the layout of FIG. 35. In addition to encroaching base portion 66A, the transistor of FIGS. 35 and 36 has an additional intrinsic base and thus an additional main intrinsic base portion 64MA. The layout of FIG. 35 is basically a combination of the layouts of FIGS. 30 and 33. Spacings S_NE and S_NEB thus both meet the requirements given above for the transistor of FIGS. 30 and 31 and the transistor of FIGS. 33 and 34.

The fabrication process of FIG. 28 or 29 can be employed to fabricate the transistor of FIGS. 30 and 31. The transistors of FIGS. 32-35 can be manufactured according to the process variation of FIG. 30 where the encroaching-base dopant is typically boron since the encroaching-base dopant is introduced into epitaxial layer 52 prior to beginning the formation of base layer 64, emitter 58, and base contact zones 68.

FIGS. 36-39 illustrate partial mask layouts respectively for the transistor layouts of FIGS. 30, 32, 33, and 35. These mask layouts are suitable for the process variation of FIG. 29.

Items

110, 110A, and 110B in FIGS. 36-39 are masks that are respectively used to define spacings S_NE, S_NEA, and S_NEB. The letters "E", "B", "C" indicate the minimum areas where the mask material must be present when etching the polysilicon layer to define the emitter, base, and collector contacts. Although not indicated in FIGS. 36-39, the polysilicon base contacts in each of the layouts typically connect to one another.

The polysilicon emitter-stripe contacts in the transistor of FIGS. 30 and 31 and in the transistor of FIGS. 34 and 35 can all be connected together. Additional such emitter stripe arrangements can be utilized to produce a power transistor. Alternatively, the emitter stripes can be left unconnected, thereby producing a multiple-emitter transistor.

While the invention has been described with reference to particular embodiments, this description is solely for the purpose of illustration and is not to be construed as limiting the scope of the invention claimed below. For example, instead of using a fast-diffusing species to create deep encroaching base portions 66 in the process of FIG. 28, encroaching base portions 66 can be formed by implanting boron deep into epitaxial layer 52 and then performing a suitably shortened anneal to drive the boron laterally outward to the desired positions for encroaching base portions 66. Semiconductor materials of opposite conductivity type to those described above can be utilized to achieve the same results as given in the above text.

The invention can be applied to high-voltage transistors fabricated on substrates having substantially higher resistivity than that of substrate 50 here. The pn-junction depths, minimum base thickness T_BMIN, and minimum non-encroached spacing S_NE are greater than in the exemplary transistors described above. Nonetheless, parameters t_BMIN and S_NE still satisfy the relationships given above. Various modifications and applications may thus be made by those skilled in the art without departing from the true scope and spirit of the invention as defined in the appended claims.

Claims

We claim:

1. A method of fabricating a bipolar transistor from a semiconductor body having a surface-adjoining major region of a first conductivity type, the major region having an upper surface, the method comprising the following steps:

(1) introducing primary dopant of a second conductivity type opposite to the first conductivity type into the major region through the major region's upper surface to define a first base layer of the second conductivity type;

(2) introducing dopant of the first conductivity type into the major region through a segment of the major region's upper surface to define an emitter of the first conductivity type such that the emitter overlies a main intrinsic base portion of the first base layer; and

(3) introducing further dopant of the second conductivity type into the major region through two laterally separated segments of the major region's upper surface, part of the further dopant being driven sideways and downward to form two encroaching base portions of the second conductivity type such that the encroaching base portions are continuous with the main intrinsic base portion, have a lighter doping than the main intrinsic base portion, extend deeper into the semiconductor body than the main intrinsic base portion, extend respectively below two corresponding parts of the main intrinsic base portion, extend laterally outward beyond the emitter, and are spaced laterally apart from each other below the main intrinsic base portion by a minimum spacing S_NE which is less than three times the minimum distance t_BMIN by which the main intrinsic base portion vertically separates the emitter and a main collector region of the first conductivity type underlying the main intrinsic base portion.

2. A method as in claim 1 wherein the encroaching base portions are driven sufficiently close to each other below the main intrinsic base portion to improve the transistor's output current-voltage characteristics.

3. A method as in claim 2 wherein the collector-to-emitter punch-through voltage is of a magnitude 20% greater than what would exist in the absence of the encroaching base portions.

4. A method as in claim 2 wherein S_NE is less than twice T_BMIN.

5. A method as in claim 2 wherein at least two of the introducing steps include diffusions that overlap time-wise.

6. A method as in claim 1 wherein the third-mentioned introducing step is initiated after the other two introducing steps are initiated.

7. A method as in claim 6 wherein, at a given temperature, the further dopant diffuses through the major region at a faster rate than the other two dopants.

8. A method as in claim 7 wherein the further dopant comprises aluminum when the first and second conductivity types respectively are n-type and p-type.

9. A method as in claim 1 wherein the third-mentioned introducing step is initiated before the second-mentioned introducing step is initiated.

10. A method as in claim 1 wherein the second-mentioned introducing step entails diffusing the dopant of the first conductivity type into the major region from an overlying non-monocrystalline semiconductor piece of the first conductivity type.

11. A method as in claim 1 wherein spacers are provided along an emitter contact provided above the emitter, and the locations of the encroaching base portions are controlled by the spacers.

12. A method as in claim 1 wherein the third mentioned introducing step entails introducing the further dopant through a masking structure having a dopant-blocking stripe which, as viewed in a direction perpendicular to the upper surface of the major region, crosses the location for the emitter, the dopant-blocking stripe substantially preventing the further dopant from passing through the dopant-blocking strip.